icon_folder.gif   Conference Reports for NATAP  
 
  10th European AIDS Conference (EACS)
Nov 17-20, 2005
Dublin, Ireland
Back grey_arrow_rt.gif
 
 
 
A Look at Lipoatrophy Predictors and Reversers
 
 
  A Report from the 10th European AIDS Conference (EACS), November 17-20, 2005, Dublin
Written for NATAP
http://www.natap.org
 
Mark Mascolini
 
TOPICS
High triglycerides foretell lipoatrophy
Drop all nukes to reverse lipoatrophy?
Factors favoring limb fat gains in RAVE
Uridine dietary supplement boosts fat

 
Two independent teams now have evidence that high triglycerides foreshadow the subcutaneous fat wasting commonly called lipoatrophy. Continuing work by other researchers suggests at least two ways to bring subcutaneous fat back—stop taking the thymidine analog nucleoside reverse transcriptase inhibitors (NRTIs) d4T and AZT, or take uridine (perhaps at some risk).
 
High triglycerides foretell lipoatrophy
At the previous European AIDS Conference (EACS) in Warsaw, the University of Milan's Massimo Galli offered evidence that high triglycerides precede lipoatrophy in people with HIV infection. At this year's Dublin EACS, he was back with more data confirming that finding [1]. Certain limitations make Galli's study less than ideal, but an independent team from Bristol-Myers Squibb reported the same link last year [2].
 
Galli's latest effort involved 839 people starting triple antiretroviral therapy from 1997 through 2001 in the ICoNA cohort. As a result most of the regimens for these people look old-fashioned by today's standards. Favored NRTI parings were AZT/3TC (936 person-years of follow-up [PYFU], d4T/3TC (423 PYFU), and d4T/ddI (255 PYFU). Unboosted indinavir (613 PYFU), nevirapine (294 PYFU), nelfinavir (169 PYFU), and efavirenz (127 PYFU) were the most frequently used third drugs. Cohort clinicians penned so few prescriptions for ritonavir-boosted protease inhibitors (PIs) that statisticians eliminated them from the analysis.
 
About three quarters of the cohort was male, 40% had HCV coinfection, 89% had a CD4 count above 200 cells/mL, and 71% took just one or two NRTIs before starting triple therapy. Lipoatrophy diagnosis rested solely on clinical assessment, not on any scanning techniques.
 
Multivariate analysis traced no link between lipoatrophy and gender, HIV transmission route, HCV coinfection, or age, although a higher risk with older age (seen in earlier studies) approached statistical significance (Table).
 
Predictors and nonpredictors of lipoatrophy in the ICoNA cohort
 

predictores-1.jpg

Among antiretrovirals tested, d4T plus ddI significantly boosted the odds of lipoatrophy in a univariate analysis, but not in multivariate calculations (Table). Only d4T plus 3TC independently upped the odds of fat atrophy (Table).
 
By itself, high cholesterol did not make lipoatrophy more likely in the ICoNA cohort, but high triglycerides did (Table). Yet the correlation between lofty triglycerides and lipoatrophy did not follow a steady “dose-response” pattern. Lipoatrophy proved significantly more likely (P = 0.02) with borderline high triglycerides (150 to 199 mg/dL), but the correlation fell just shy of statistical significance (P = 0.07) with high triglycerides (200 to 499 mg/dL). Very high triglycerides (500 mg/dL or more) proved the highest independent predictor of lipoatrophy.
 
Besides reliance on physician-diagnosed lipoatrophy and use of now-outmoded regimens, Galli noted one other limitation to his analysis. It could not weigh the potential influence of diet, behavior, or other factors on high triglycerides and lipoatrophy.
 
Yet in a similarly large cohort of d4T trial enrollees, Bristol-Myers Squibb workers also independently correlated high pretreatment triglycerides with physician-defined lipoatrophy [2]. In a report from early 2004, Mustafa Noor and colleagues looked for lipoatrophy predictors in 877 people who took either extended-release d4T or immediate-release d4T with 3TC and efavirenz in two trials. Median follow-up exceeded 2 years. Unlike Galli, they looked at only two pretreatment triglyceride brackets—fewer or more than 200 mg/dL. In people taking extended-release d4T, baseline triglycerides had no impact on new diagnoses of lipoatrophy. But in the immediate-release d4T group, 46 of 382 (12%) with pretreatment fasting triglycerides under 200 mg/dL had lipoatrophy, compared with 16 of 59 (27%) who started immediate-release d4T with 200 mg/dL of triglycerides or more.
 
Multivariate analysis tied pretreatment fasting triglycerides under 200 mg/dL, age under 40 years, and extended-release d4T to a lower risk of fat atrophy.
 
Drop all nukes to reverse lipoatrophy?
Several studies now confirm that replacing d4T or AZT with abacavir or tenofovir slowly but certainly reverses limb fat loss. One of these trials, presented at the Lipodystrophy Workshop in Dublin just before EACS, found that swapping d4T or AZT for either of those drugs also restores lost facial fat [3]. After 48 weeks on abacavir or tenofovir, cheek fat measured by 3D laser scans averaged 2539 mm3, about as much as one sees with collagen injections. Yet only about a third the people in this trial noticed the 48-week improvement.
 
If dumping a thymidine analog (d4T or AZT) and starting abacavir or tenofovir restores limb and face fat, would it help to avoid all NRTIs? That's the question asked by Marc-Antoine Valantin from the Pitie-Salpetriere Hospital in Paris and colleagues around France in the NO-NUKE trial (ANRS protocol 108) [4].
 
The study started with 100 people whose viral load lay below 400 copies/mL with their NRTI-containing regimen. Valantin randomized them to continue that regimen or to switch to a ritonavir-boosted PI (usually indinavir or lopinavir) plus a nonnucleoside. The study groups had similar NRTI experience (6.6 years in the nuke arm and 6.4 years in the no-nuke arm), and most people were taking d4T or AZT before randomization (68% in the nuke arm and 74% in the no-nuke arm). Almost everyone had tried d4T at some point.
 
Forty-eight weeks later an intent-to-treat analysis discerned no virologic differences between arms. Nearly everyone kept their viral load under 400 copies/mL. But subcutaneous thigh fat measured by CT scan in a central lab by radiologists blinded to treatment assignment found an average 34 cm3 fat gain in the no-nuke group versus no change in the nuke-maintenance group (P = 0.002).
 
Stopping d4T or AZT explained almost all of the thigh fat gains. Among 29 people not taking d4T or AZT at baseline, thigh fat meandered nonsignificantly upward in both the nuke group and the no-nuke group. But among 71 people taking d4T or AZT at enrollment, the no-nuke group added an average 43.7 cm3 in 48 weeks while the nuke group lost an average 8.5 cm3 (P = 0.001).
 
Even among people stopping AZT or d4T, observed Andrew Carr from Sydney's St. Vincent's Hospital, these are not big subcutaneous fat gains. Mean baseline thigh fat measured 299 cm3, so a 43.7 cm3 mosey after stopping d4T or AZT represents only a 14.6% uptick and brings people nowhere near normal thigh fat values.
 
And although switching to a PI/nonnucleoside regimen did not jeopardize viral suppression in this study, it did jack up triglycerides by 1.32 mmol/L (versus a 0.13 mmol/L drop in the NRTI maintenance arm, P = 0.002). On the other hand, aspartate aminotransferase (AST) dropped in the no-nuke group (-8.1 U/L) while rising in the nuke group (+3.9 U/L) (P = 0.008).
 
But the key question remains whether lipoatrophy justifies swapping a standard triple regimen for an unorthodox NRTI-sparing mix. After Valantin's report, Jose Gatell from Barcelona's Hospital Clinic suggested that PI/nonnucleoside regimens have not yielded reliable virologic response rates. And another lipoatrophy trial by French researchers had to end early for precisely that reason.
 
The Hippocampe study randomized treatment-naive people to a standard NRTI-containing triple regimen or to a nonnucleoside plus a boosted PI to see if lipoatrophy would appear faster with the NRTIs [5]. But they never found out. The Data Safety and Monitoring Board shut down the study early because the standard triple regimen slowed HIV significantly better than the two-class combo.
 
With colleagues throughout France, Claudine Duvivier from Paris's Pitie-Salpetriere Hospital randomized 117 untreated people in a 2:1:1 ratio to a boosted PI plus a nonnucleoside, a boosted PI plus two NRTIs, or a nonnucleoside plus two NRTIs. Clinicians had the option of using lopinavir/ritonavir or indinavir/ritonavir as the PIs, efavirenz or nevirapine as the nonnuke, and any NRTI except d4T or ddC.
 
Everyone began treatment with a CD4 count under 350 cells/mL and a viral load above 5000 copies/mL. Median pretreatment loads measured 5.0 log copies/mL in the PI/nonnucleoside arm and 5.8 log copies/mL in the NRTI arms. Respective median baseline CD4 counts stood at 214 and 203 cells/mL.
 
For the virologic analysis, Duvivier grouped the two NRTI-containing arms and compared them with the NRTI-sparing arm. In the NRTI-sparing group, 75% took lopinavir/ritonavir, 25% took indinavir/ritonavir, 63% took efavirenz, and 37% took nevirapine. In the NRTI/PI group, 52% took lopinavir/ritonavir and 48% indinavir/ritonavir. And in the NRTI/nonnucleoside group, 79% took efavirenz and 21% nevirapine.
 
Even though people taking an NRTI-containing regimen started treatment with nearly a 10-fold higher viral load than those taking an NRTI-sparing regimen, virologic response proved significantly better in the NRTI-containing arms at several early points by either a missing-or-change-equals-failure analysis or an on-treatment analysis (Table).
 
Response to NRTI-sparing therapy versus standard NRTI-containing triple regimens
 

week4-2.jpg

CD4 gains didn't differ in the NRTI and no-NRTI arms through 48 weeks.
 
These results appear to bolster a familiar rule of antiretroviral prescribing: Don't mess around with offbeat regimens unverified in randomized trials. But two attendees argued that this trial does not prove NRTI-sparing regimens lack the punch of standard NRTI-based medleys for first-line therapy.
 
First, one attendee observed, the study's primary endpoint was lipoatrophy after 96 weeks, so statisticians hadn't powered the trial to weigh virologic response. But Duvivier reported that the 36-week virologic analysis was robust enough (in tongue-tangling statistical parlance) to reject the premise that the NRTI-sparing regimen is noninferior to the NRTI-backbone regimen.
 
Second, Daniel Kuritzkes (Brigham and Women's Hospital, Boston) asked whether the French team had analyzed PI plasma levels, noting that efavirenz can cut PI concentrations and that lower PI levels would jeopardize response. Indeed, responded Hippocampe investigator Christine Katlama, a preliminary pharmacokinetic analysis saw a link between lower PI levels and virologic failure.
 
But PI and/or nonnucleoside concentrations climbed high enough to cause side effects. Difficulty stomaching the boosted PI/nonnucleoside pairing undermined that regimen in the missing-or-change analysis. While only 1 person (2%) taking an NRTI-based regimen had to cash in one drug for another, 11 people (18%) in the NRTI-sparing group had to change drugs.
 
While awaiting more detailed analyses of these two trials, the prudent clinician might conclude that dumping all NRTIs—rather than just d4T or AZT—looks like an overly aggressive way to ward off or reverse lipoatrophy.
 
Factors favoring limb fat gains in RAVE
RAVE is one of the several randomized trials that verified the value of replacing d4T or AZT with a nonthymidine nucleoside like abacavir or tenofovir. In Dublin the RAVE team, headed by Graeme Moyle of London's Chelsea and Westminster Hospital, offered further analysis of that trial showing that people switching from AZT (versus d4T), older people, and white people gained less limb fat after 48 weeks of tenofovir or abacavir [6].
 
Measuring limb fat with DEXA scans, Moyle reported that median limb fat climbed 393 grams with tenofovir in 52 people and 316 grams with abacavir in another 52, a nonsignificant difference. But among people who replaced AZT in their regimen, median 48-week limb fat gains measured only 66 grams with tenofovir and 210 grams with abacavir. In contrast, median gains among people replacing d4T measured 529 grams in the tenofovir group and 357 grams in the abacavir group.
 
Moyle hatched two multivariate analyses to sort out independent predictors of fat bouncebacks in the trial. An analysis focused on absolute limb fat change zeroed in on two predictors:
 
- Every extra 5 years of age an randomization meant 103 fewer grams of recovered fat at week 48 (95% confidence interval 214 gram less to 8 grams more, P = 0.07). - Replacing AZT with tenofovir or abacavir meant 419 fewer grams of recovered fat at week 48 (95% confidence interval 65 to 774 grams less, P = 0.02).
 
A multivariate model reckoning odds for more than a 710-gram limb fat jump at 48 weeks rooted out three predictors:
 
- Replacing AZT with tenofovir or abacavir lowered the chance of gaining 710 grams 85% (odds ratio 0.15, 95% confidence interval 0.03 to 0.66, P = 0.01). - Every extra year of antiretroviral therapy before switching to tenofovir or abacavir lowered the chance of gaining 710 grams by 25% (odds ratio 0.75 per year, 95% confidence interval 0.57 to 0.98, P = 0.04). - Being white lowered the chance of gaining 710 grams 83% (odds ratio 0.17, 95% confidence interval 0.04 to 0.69, P = 0.01).
 
Moyle suggested that switching from d4T or AZT earlier in the course of treatment will improve chances of gaining limb fat.
 
Uridine dietary supplement boosts fat
NucleomaxX, a uridine-based dietary supplement on the shelves of European health food stores, revived lost limb fat in people taking d4T or AZT [7]. But this pilot study was small and short, the remedy may work only for people taking d4T or AZT, and people taking uridine gained pernicious visceral adipose tissue while their “good” high-density lipoprotein (HDL) cholesterol dropped. Two other things—NucleomaxX tastes awful and isn't cheap.
 
This 3-month double-blind, placebo-controlled trial by Helsinki University's Jussi Sutinen rests on cell-based studies by Frieburg University's Ulrich Walker, who collaborated with Sutinen and holds a patent on using uridine for fat atrophy. Walker's work showed that uridine staved off d4T- or AZT-induced mitochondrial depletion in fat cells [8].
 
Sutinen signed up 20 adults taking an effective d4T- or AZT-containing regimen for at least 1.5 years and randomized them to uridine or placebo. The trial excluded anyone taking ddI (because uridine doesn't work on ddI-spawned mitochondrial mix-ups) and anyone allergic to milk protein, a NucleomaxX ingredient.
 
For the first 10 days of each study month, people randomized to uridine took 36 grams three times daily. One person who couldn't stand the taste (which was approximated in the placebo powder) dropped out of the study, and one person in the placebo arm died of a heart attack. Of the 9 people who took uridine for 3 months, five were taking an AZT regimen and four a d4T regimen. No one switched or stopped antiretrovirals during the trial.
 
After 3 months average limb fat leapt from a pretreatment reading of 3370 grams (+ 890 grams standard error of mean, SEM) to 4260 grams (+ 940 grams SEM), a significant gain (P < 0.05 versus baseline or placebo). Sutinen also saw significant jumps in total body fat (P < 0.01) and portentous intra-abdominal fat (P < 0.05) with uridine. Average intra-abdominal fat climbed from 2320 cm3 (+ 320 cm3 SEM) at baseline to 2520 cm3 (+ 370 cm3 SEM) with 3 months of uridine (P < 0.05) but did not rise among people taking placebo. Uridine had no significant impact on liver fat.
 
In the uridine group the proportion of limb fat to total fat flared from 19% to 25% (P < 0.05). Everyone taking uridine added limb fat, but no one noticed the change.
 
A slightly longer uridine study with more intense dosing found that both patients and physicians noticed improvements in lipoatrophy and lipohypertrophy (fat buildups), but objective fat measures failed to confirm those impressions [9]. Unlike Sutinen's trial, this study by Grace McComsey at Case Western Reserve University in Cleveland was not blinded or placebo controlled.
 
McComsey recruited 16 people with lipoatrophy while taking a d4T regimen. Continuing their antiretrovirals, study participants took 36 grams of NucleomaxX three times daily every other day for 16 weeks. Follow-up without NucleomaxX lasted another 16 weeks. McComsey calculated adherence at better than 95% in all but one person.
 
Fourteen people completed the trial. They believed they saw significant improvements in lipoatrophy, reporting an average body-area-based score of 7.4 before NucleomaxX, 5.3 after 16 weeks of treatment, and 5.0 after a further 16 weeks of follow-up without treatment (P < 0.05). (A lower score means less lipoatrophy.) Scores calculated by their physicians for the same three points were 8.6, 5.7, and 5.9 (P < 0.05). Both physicians and study participants also registered significant improvements in lipohypertrophy at weeks 16 and 32.
 
McComsey did not measure limb or visceral fat changes, but body mass index varied hardly at all during the study, starting at an average 24.3 kg/m2, then drifting to 24.8 kg/m2 at week 16 and to 24.6 kg/m2 at week 32. Mitochondrial DNA content never changed in fat cells or peripheral blood mononuclear cells.
 
Sixteen weeks of uridine yielded no significant shifts in total cholesterol, non-HDL cholesterol, lactates, glucose, insulin, insulin resistance measured by HOMA, CD4 count, or viral load. Triglycerides climbed nonsignificantly with NucleomaxX, from an average 154 mg/dL at study entry to 181 mg/dL at week 16. After another 16 weeks of follow-up without the uridine-rich powder, average triglycerides retreated to 140 mg/dL.
 
In Sutinen's study average HDL cholesterol inched upward in the placebo arm but dove from 1.24 + 0.1 mmol/L at enrollment to 1.15 + 0.08 mmol/L after 3 months of uridine (P < 0.05). Uridine did not significantly affect insulin resistance, triglycerides, alanine aminotransferase, venous blood pH, CD4 count, or HIV load during the 3-month trial.
 
In an EACS review of lipodystrophy news, Andrew Carr estimated that NucleomaxX costs 250 Euros monthly. But Sutinen figured a 1-month supply for the tested dose would set users back 84 Euros if purchased over the Internet. The trial will continue for another 9 months to see if a lower, cheaper dose can maintain or improve fat gains. Everyone will also want to watch long-term effects on visceral fat, HDL cholesterol, triglycerides, and other measures that may change with continued use.
 
Mark Mascolini writes about HIV infection
 
References
1. Galli M, Cozzi-Lepri A, Ridolfo AL, et al. Triglyceride level predicts lipoatrophy in patients receiving antiretroviral therapy. 10th European AIDS Conference. November 17-20, 2005. Dublin. Abstract PS5/1.
 
2. Noor M, Dezii C, McLaren C, et al. Baseline triglyceride levels predict development of lipoatrophy in patients treated with stavudine extended-release/prolonged-release capsules or stavudine immediate-release. 11th Conference on Retroviruses and Opportunistic Infections. February 8-11, 2004. San Francisco. Abstract 722.
 
3. Benn P, Sauret V, Cartledge J, et al. Improvements in facial lipoatrophy at 48 weeks following substitution of a thymidine analogue with tenofovir or abacavir: a randomized, open-label study in people with lipoatrophy and virological suppression on HAART. 7th International Workshop on Adverse Drug Reactions and Lipodystrophy in HIV. November 13-16, 2005. Dublin. Abstract 8.
 
4. Valantin MA, Lanoy E, Bentata M, et al. Maintenance of virological suppression and impact on lipoatrophy of NRTI-sparing regimens in antiretroviral-treated patients: NONNUKE ANRS 108 study. 10th European AIDS Conference. November 17-20, 2005. Dublin. Abstract 5/2.
 
5. Duvivier C, Ghosn J, Assoumou L, et al. Lower rate of virological suppression in naive patients initiating HAART with NRTI-sparing regimen compared to standard NRTI-containing regimen: results from HIPPOCAMPE—ANRS 121 trial. 10th European AIDS Conference. November 17-20, 2005. Dublin. Abstract PS1/3.
 
6. Moyle G, Sabin C, Cartledge J, et al. Factors associated with limb fat recovery in a prospective randomized comparative study of thymidine replacement with either tenofovir DF or abacavir in persons with clinical lipoatrophy. 10th European AIDS Conference. November 17-20, 2005. Dublin. Abstract PE9.3/2.
 
7. Sutinen J, Walker UA, Häkkinen AM, et al. Uridine for the treatment of HAART-associated lipodystrophy: a randomized, double-blind, placebo-controlled trial. 10th European AIDS Conference. November 17-20, 2005. Dublin. Abstract PS5/6.
 
8. Walker UA, Venhoff N. Uridine in the prevention and treatment of NRTI-related mitochondrial toxicity. Antivir Ther 2005;10(suppl 2):M117-M123.
 
9. McComsey GA, Setzer B, O'Riordan M, et al. NucleomaxX improves clinical lipoatrophy without any change in fat mitochondrial DNA levels. 10th European AIDS Conference. November 17-20, 2005. Dublin. Abstract LBPE9.4/8.